Forum on Dyslexia : Dyslexia as a Phonological Deficit

Dyslexia as a Phonological
Deficit : Evidence and
Implications
Margaret Snowling

In recent years, limitations of the discrepancy definition of specific reading

difficulties have led researchers to propose that dyslexia is best described
as a core phonological deficit. With studies of the normal development of
reading as a backdrop, this paper reviews the phonological processing
deficit hypothesis of dyslexia. The extant evidence suggests that
phonological difficulties in dyslexia persist throughout development from
the preschool years into adulthood, despite compensation of reading
deficits. Moreover, individual differences in the behavioural manifestation
of dyslexia are consistent with differences in the severity of underlying
phonological difficulties. Implications for the early identification of
dyslexia and for interventions to prevent reading failure are discussed.

Keywords : Dyslexia ; phonological awareness ; reading difficulties ; phonological disability

Introduction Defining dyslexia

Developmental dyslexia was first described just over 100 The regression approach to dyslexia
years ago in the British Medical Journal (Morgan, 1896). Since the landmark Isle of Wight studies, it has become
That case report described a 14-year-old boy named Percy widely accepted in educational and clinical practice that it
who, in spite of being of normal intelligence, had been is important to distinguish children who have specific
unable to learn to read. Extracts of Percy’s writing reveal reading difficulties (specific reading retardation) from
many characteristic signs of dyslexia including dys- children who have reading difficulties in the context of
phonetic spelling errors (carefully calfuly), substi- more general learning problems (reading backwardness)
tutions of phonemes (peg pag) and transpositions of (Rutter & Yule, 1975 ; Silva, McGee, & Williams, 1985). It
letters (Percy Precy). Such difficulties are now thought is the former group, those with specific reading difficulties,
to arise because dyslexic people have poor phonological to whom the term ‘ dyslexia ’ is usually applied. The
skills that prevent them learning the rule-governed system definition of specific reading difficulties depends upon the
of English spelling. However, in spite of advances in use of a regression approach ; a child is deemed to have a
understanding of both the cognitive processes involved in specific problem with reading if, and only if, their reading
the acquisition of literacy and of how difficulties with this attainment is significantly below that predicted from their
process arise, there is continuing debate surrounding the general cognitive ability on the basis of the correlation
use of the term ‘ dyslexia ’. In this article I will propose that between reading and IQ in children of the same age. In
it is possible to arrive at a coherent definition of dyslexia other words, if they show a discrepancy between expected
that is theoretically motivated and of clinical utility. To do and actual attainment.
Margaret Snowling so it is necessary to adopt a developmental perspective in
Department of which the behavioural manifestations of core deficits The prevalence of specific reading difficulties (dyslexia), as
Psychology, University change over time as a consequence of developmental determined by the regression approach, depends critically
of York, York YO1 5DD interactions and compensatory processes. upon the cut-off point taken as indicative of reading

Child Psychology & Psychiatry Review Volume 3, No. 1, 1998

4
 Forum on Dyslexia : Dyslexia as a Phonological Deficit

disability. Assuming a normal distribution of reading reading difficulty at any one time, equivalent in practical
achievement scores, 2.28 % of children should score more terms to a retardation of some 24 months in reading
than two standard errors of measurement below their relative to expectation.
expected attainment score. Using this cut-off, Yule et al.
(1974) reported a prevalence of specific reading retar- Thus, it is important to recognise limitations inherent in
dation ranging from 3.1 % among 10-year-olds in the Isle the regression approach to the diagnosis of dyslexia. At a
of Wight to 6.3 % in London, based on the discrepancy fundamental level, Stanovich (1986a) has highlighted the
between IQ and reading accuracy. The comparable figures problems surrounding the assessment of IQ in poor
for reading comprehension were 3.6 % and 9.3 %. These readers whose verbal skills may decline as a consequence
figures illustrate clearly that prevalence depends both of their limited reading experience ; with no diagnostic
upon geographical distribution (figures were higher in an criteria to fulfil other than statistical ones, the procedure
inner-city population) and also on the tests used to runs the risk of identifying false positives and false
determine both IQ and reading skill. A national survey in negatives. In practice, little is known about the charac-
Britain using a reading test with better psychometric teristics of so-called false positives, who may include
properties reported a lower prevalence of 2.29 % children whose reading problems arise alongside emo-
(Rodgers, 1983) and a similar study in New Zealand found tional or behavioural difficulties. Such children would
only 1.2 % of 9-year-olds to have specific reading re- normally be excluded from empirical investigations, al-
tardation (Silva et al., 1985). though clinically, their reading problems warrant further
investigation and treatment. From an educational per-
The most recent epidemiological data from a longitudinal spective, false negatives produce more contention ; these
study of 414 Connecticut children used a less stringent cut- children, as a consequence of remediation or of compen-
off of 1.5 standard errors of measurement below ex- sation, have overcome their primary reading problems
pectation and reported prevalence rates of 5.6 % in first and no longer fulfil the criteria for specific reading
grade, 7 % in third grade, and 5.4 % in fifth grade (B. A. retardation but may have serious problems with spelling
Shaywitz et al., 1992). Although a reservation is that the and written work.
numbers of children fulfilling the definition of specific
reading retardation were small in this sample, it is salutary In summary, the regression methodology may particularly
to note that the diagnosis of dyslexia, defined by the disadvantage the children whom it is meant to identify,
regression approach, was not a stable entity. Rather, there and it is noteworthy that the discrepancy definition is
was predictable year-to-year variability ; only 28 % of silent with regard to the identification of children at risk of
children defined as dyslexic at the end of first grade were reading failure. For clinical utility, the regression ap-
given the same diagnosis by third grade and only 47 % of proach needs to be supplemented by positive diagnostic
these fulfilled the criteria for specific reading retardation features that will allow the practitioner to identify children
in fifth grade (cf. McGee et al., 1986). who show early or residual signs of dyslexia that require
intervention and do not depend solely on the extent of the
Notwithstanding the difficulties in arriving at precise child’s failure relative to the norm.
figures for the prevalence of specific reading retardation,
studies have produced conflicting results in regard to its Dyslexia as a core phonological deficit (CPD)
distribution between the sexes. The earlier epidemiological If discrepancy-based definitions of dyslexia are to be
studies typically reported an excess of males with specific supplemented, then what should the new definition
reading difficulties (about 4 boys to 1 girl), consistent with comprise ? Clearly the answer to this question depends
the findings from referred samples. However, in their upon the theoretical framework that one adopts. Given
epidemiological sample S. E. Shaywitz et al. (1990) found that dyslexia is a developmental disorder, it would be
a more even distribution of between 1.3 and 1.5 boys to advantageous to agree on a definition that would embrace
every girl affected. Interestingly, within the sample who the differences in symptomatology at different points in
had been referred by their schools as having specific development. Similarly, the definition should be capable
reading difficulties, there was a bias towards boys (a of encompassing the surface manifestations of dyslexia in
gender ratio of 4.3 : 1 in second grade and 2.4 : 1 in third different cultures, languages, and school systems.
grade). Children who were identified by their schools as
having reading difficulties but who did not fulfil the Although variants of a visual deficit hypothesis of dyslexia
regression-based definition were more likely to exhibit are still in evidence (see Eden et al., 1996 ; Lovegrove &
behaviour problems than non-referred children who Williams, 1993), the most widely accepted view is that
nonetheless had specific reading difficulties, and there was dyslexia can be considered part of the continuum of
a tendency for girls to have more severe reading problems language disorders and is a verbal processing deficit
before they were referred. Thus, patterns of referral as well (Vellutino, 1979). Since 1980, a large number of studies
as differences in teaching practice may affect prevalence have appeared in the literature pointing to phonological
rates for ‘ dyslexia ’ in different samples. Taking account of processing difficulties in dyslexic children (Shankweiler &
the available evidence, it seems reasonable to infer that Crain, 1986 ; Share, 1995 ; Snowling, 1995, for reviews).
between 4 % and 7 % of children will have a specific The most consistently reported of these difficulties are

Child Psychology & Psychiatry Review Volume 3, No. 1, 1998

5
Forum on Dyslexia : Dyslexia as a Phonological Deficit

problems with phonological awareness and limitations of However, there are also dyslexic children who appear to
verbal short-term memory. There is also evidence that have mastered alphabetic skills. Such children have been
dyslexic children have trouble with long-term verbal referred to as developmental ‘ surface ’ or ‘ morphemic ’
learning. This problem may account for many classroom dyslexics (Coltheart et al., 1983 ; Seymour, 1986). The
difficulties, including problems memorising the days of the classic characteristic of these children is that, in single-
week or the months of the year, mastering multiplication word reading, they rely heavily upon a phonological
tables, and learning a foreign language. strategy. Thus, they tend to pronounce irregular words as
though they were regular (e.g. glove gloave ; island
Dyslexic children also have difficulties with the retrieval of izland), they have particular difficulty distinguishing
phonological information from long-term memory. between homophones like pear–pair and leek–leak, and
Word-finding difficulties are often seen clinically and their spelling is usually phonetic.
experimental studies using both rapid naming and object
naming tasks report deficiencies. It is possible that these Although evidence in favour of distinct subtypes is lacking
problems stem from more basic deficits in speech per- (cf. Bryant & Impey, 1986 ; Seymour, 1986), most sys-
ception, speech production, or temporal processing tematic studies of individual differences among dyslexics
(Farmer & Klein, 1995), and a number of current theories have revealed variations in their reading skills (Castles &
suggest that dyslexic children have difficulties first in Coltheart, 1993). A number of recent studies have
establishing, and later in accessing, adequate phonological suggested that this variation might be explained by
representations (Fowler, 1991 ; Hulme & Snowling, 1992 ; differences in the severity or developmental status of
Swan & Goswami, 1997). Such difficulties may ultimately underlying phonological skills. One way of characterising
explain more wide-ranging cognitive deficits in dyslexic a child’s reading is to assess how well they can decode
children (Snowling & Hulme, 1994). words they have not seen before, for example, using a
nonword reading test, and how well they recognise words
The strength of the evidence pointing to the phonological that they cannot ‘ sound out ’, such as irregular or
deficits associated with dyslexia led Stanovich (1986b) to exception words that do not conform to English spelling
propose that dyslexia should be defined as a core rules (e.g. pint, yacht). Manis and colleagues (1996) and
phonological deficit. Importantly, within the phonological also Stanovich, Siegel, and Gottardo (1997) found that
core-variable difference model of dyslexia (Stanovich & dyslexic children who had relatively more difficulty in
Siegel, 1994), poor phonology is related to poor reading reading nonwords than exception words (phonological
performance, irrespective of IQ. Skills close to the ‘ core ’ dyslexia) performed significantly less well than reading-
of dyslexia include nonword reading and aspects of age-matched younger controls on tests of phonological
phonological awareness ; all poor readers will tend to awareness. In contrast, dyslexic children who had more
differ from normal readers in these skills. Children with difficulty with exception words than nonwords (classified
specific reading difficulties (discrepancy-defined poor as surface dyslexic) performed at a similar level to controls
readers) differ from generally poor readers (garden-variety on these tests. Both groups of researchers argued for a
poor readers) in skills further from the core, including similar interpretation of their results ; they proposed that
measures of working memory and listening comprehen- children with phonological dyslexia have a disorder of
sion (see also B. A. Shaywitz et al., 1992). development whereas surface dyslexics are develop-
mentally delayed.
An important advantage of the core phonological deficit
definition of dyslexia is that it makes sense in terms of In a series of single-case studies from our laboratory, we
what is known about the normal acquisition of reading. have also compared the performance of children who
Since the pioneering studies of Bradley and Bryant (1983) differ in reading profile on tests of phonological pro-
in Britain, and Lundberg and colleagues in Denmark cessing. Generally, children with more severe phono-
(Lundberg, 1994), it has been known that phonological logical processing impairments show more significant
awareness measured in preschool is an excellent predictor impairments in their decoding skills and in their ability to
of subsequent reading achievement, even when the sub- spell phonetically (Snowling, Goulandris, & Stackhouse,
stantial effects of IQ are controlled. Indeed, the ability to 1994). Furthermore, Snowling, Goulandris, and Defty (in
reflect upon the sound structure of words at the phonemic press) followed two children who displayed a ‘ phono-
level is critical to the development of the alphabetic logical dyslexic ’ style of reading (in reading words
principle that allows children to decode words they have significantly better than nonwords), and two children who
not seen before (Byrne & Fielding-Barnsley, 1989). Dys- resembled ‘ surface dyslexics ’ (in reading irregular words
lexic children typically fail to make the transition from significantly less well then regular words) over a 2-year
visual or logographic reading to the alphabetic phase period. At both points in time, the two children who
(Frith, 1985). They continue for longer than normal showed a ‘ phonological dyslexic ’ profile had more diffi-
readers to rely on a visual strategy for reading words and culty with phonological processing tasks, including tests
many have persisting difficulties reading novel words that of rhyme, nonword repetition, and phonetic spelling than
are not part of their sight vocabulary, such as nonwords the two who showed a ‘ surface dyslexic ’ profile though
(Rack, Snowling, & Olson, 1992). they did not differ in visual processing skill. It is important

Child Psychology & Psychiatry Review Volume 3, No. 1, 1998

6
 Forum on Dyslexia : Dyslexia as a Phonological Deficit

to emphasise, however, that the ‘ surface ’ dyslexics did not study in which at-risk children were worse at repeating
perform within the normal range for their age on the novel words, especially those with phonological structures
phonological tasks ; in line with findings of Manis et al. comprising late-acquired forms at 45 months. In addition,
(1996) and Stanovich et al. (1997) they appeared to show their knowledge of nursery rhymes, a test known to
a lag in their development. predict reading achievement, was poorer and, strikingly,
they already knew fewer letters.
Taken together, these findings are compatible with the
hypothesis that dyslexic reading difficulties stem from Thus, there are strong reasons to suggest that children at
phonological processing problems. However, they suggest risk of dyslexia are delayed in their phonological de-
that the severity of a child’s phonological difficulty can velopment. A reasonable hypothesis is that these children
affect the way in which their reading system becomes set come to the task of learning to read with poorly specified
up—and whether they look like ‘ phonological ’ or ‘ sur- phonological representations and that it is immaturity in
face ’ dyslexics. A second factor that may interact to these skills that compromises literacy development by
influence the reading style a child adopts is the type of placing limitations on their ability to establish the map-
teaching they receive. pings between letter strings and phonology that are critical
for learning to read.
Developmental changes in dyslexia
A second advantage of the core phonological deficit Biological evidence in support of the core phonological
hypothesis of dyslexia is that it accounts well for the deficit hypothesis
different manifestations of dyslexia seen across the life- It has been known for many years that poor reading tends
span. Although reading problems tend to be the key to run in families and there is now conclusive evidence that
behavioural symptom of dyslexia in the early school years, dyslexia is heritable (Pennington, 1994). Behaviour gene-
many adults with a childhood history of dyslexia are fluent ticists have shown that there is as much as a 50 %
readers though few overcome their spelling problems. A probability of a boy becoming dyslexic if his father is
number of recent studies have shown that adult dyslexics dyslexic (about 40 % if his mother is affected), and a
have difficulty decoding words they have not seen before somewhat lower probability of a girl developing dyslexia.
(i.e. nonwords) and have difficulties with phonological What is inherited is not, of course, reading disability per
awareness, speeded naming, and verbal short-term mem- se, but aspects of language processing. Results of large-
ory tasks (Bruck, 1990, 1992 ; Pennington et al., 1990 ; scale twin studies suggest there is greater heritability of
Snowling et al., 1997). phonological (‘ phonic ’) than visual aspects of reading. In
turn, phonological reading skills share heritable variance
At the other end of the age spectrum, there are now a with phonological awareness, the ability to reflect upon
number of family studies of dyslexia investigating differ- the sound structure of spoken words (Olson et al., 1989).
ences between children from dyslexic and non-dyslexic
families before they learn to read. Scarborough (1990) Recently studies using brain-imaging techniques to exam-
followed the development of children aged 2 to 7 years ine functional processing differences between dyslexic and
who were ‘ at risk ’ of dyslexia by virtue of having one normal brains are beginning to elucidate possible reasons
dyslexic parent. When the children were aged 7 years and why dyslexics have difficulty with phonological pro-
their reading skills could be assessed, it was possible to cessing. In one such study, Paulesu and colleagues (1996)
compare retrospectively the preschool data of children investigated differences in brain function between dyslexic
who went on to become dyslexic with children who did not and normal readers whilst they performed two phono-
develop reading difficulties. An important difference logical processing tasks. This study involved five young
between the groups was in their early language skills. adults with a well-documented history of dyslexia ; all of
Although the dyslexic children used as large a range of these dyslexics had overcome their reading difficulties but
vocabulary as their non-dyslexic counterparts at 2" years, they had residual problems with phonological awareness.
#
they made more speech errors and their use of syntax was Under PET scan, they completed two sets of parallel
more limited. At 3 years, the dyslexic children had more tasks. The phonological tasks were a rhyme judgement
difficulty with object naming and at 5 years, their and a verbal short-term memory task ; the visual tasks
difficulties extended to problems with phonological aware- were visual similarity judgement and visual short-term
ness. Their emerging literacy skills were also poorer ; they memory.
were less familiar with the letters of the alphabet and
worse at matching pictures with print. Although these dyslexic adults had compensated well for
their developmental difficulties, they showed different
Very similar results are reported by Byrne et al. (in press), patterns of left-hemisphere brain activation from controls
who found that problems of phonological awareness, during performance on the phonological processing tasks.
together with limitations of letter knowledge, presaged Specifically, there was significantly reduced activity in the
reading difficulties in high-risk children at 55 months (see insula, a region of the brain that is located between
also Lefly & Pennington, 1996). Gallagher, Frith, and Broca’s and Wernicke’s area. This area is known to be
Snowling (1996) reported converging findings from a involved in the transmission of language and, plausibly,

Child Psychology & Psychiatry Review Volume 3, No. 1, 1998

7
Forum on Dyslexia : Dyslexia as a Phonological Deficit

allows the translation between spoken inputs and speech are causally linked with dyslexic children’s reading impair-
production. It is therefore possible to speculate that the ments, it is a distinct possibility that visual problems will
insula is precisely where a ‘ disconnection ’ would be affect reading acquisition, perhaps most especially in
observed, at the cognitive level, as a difficulty in setting up children who have phonological difficulties and need to
phonological representations. rely on visual skills to compensate. Similarly, some
dyslexic children have motor impairments and others
Limitations of the core phonological deficit definition problems of attention control (Pennington, Grossier, &
In spite of the attractions of the core phonological Welsh, 1993). There is as yet very little research examining
definition of dyslexia, there are at present a number of the interactions among these different disorders. However,
outstanding issues that must be resolved before the it might be that these interactions contribute to the
definition can be fully operationalised. The first, and from heterogeneity observed among dyslexic readers in their
a practitioner’s point of view the most important, is literacy skills.
knowing what level of phonological processing skill to
expect of a child and at what age. Related is the issue of
what is the ‘ gold standard ’ test of a phonological deficit. Conclusions and clinical implications
Although some phonological processing tasks might be The argument that has been proposed here is that dyslexia
considered ‘ pure ’ and performance on them will be should be considered a core phonological deficit. Ac-
relatively independent of IQ (e.g. nonword repetition), cording to this view, dyslexic individuals of all ages
complex metalinguistic tasks tapping phonological aware- display phonological processing problems. However, the
ness are more highly correlated with IQ and naming tasks course of reading development followed by a particular
more closely tied to vocabulary development. With regard dyslexic child will be determined not only by the severity
to these latter tasks at least, the whole-scale abandonment of their phonological processing problem, but also by
of general cognitive ability in the diagnosis of dyslexia their other language skills. Dyslexic children who have
would be hazardous. good semantic skills and poor phonology are likely to
follow an atypical course of development. Their word
From a theoretical point of view, a more important recognition may improve in the face of decoding deficits
limitation of the core phonological definition of dyslexia is because they rely heavily on context. In contrast, dyslexic
its focus on phonology and the development of decoding children who have milder phonological difficulties or
skills. Although there is a consensus that automatic poorer general language skills will tend to show delayed
decoding skills are critical to reading development, Gough development along normal lines. Furthermore, the en-
and Tunmer (1986), among others, have shown that vironmental variables impinging upon the dyslexic child
proficient reading is the product of both decoding and are critical to eventual outcome. Those who receive
linguistic comprehension skills. It is important to take appropriate early intervention could quite well avoid
account of this model of reading when considering the reading failure, and cross-linguistic studies are beginning
development of dyslexic children. Although dyslexic to reveal that the manifestations of dyslexia depend upon
readers have basic level decoding difficulties, many have the orthography in which they are learning to read
normal linguistic comprehension. Typically this affords (Goswami, 1997 ; Wimmer, 1996).
them a number of advantages ; they tend to comprehend
what they read relatively well and they can monitor their The convergence of evidence from studies of the normal
comprehension processes. In some recent work, we have development of reading and from dyslexia places prac-
shown that dyslexic children show larger effects of context titioners in an excellent position both to identify children
on single-word reading than younger normal readers who at risk of dyslexia and to provide early intervention.
decode at the same level (Nation & Snowling, in press). To Children who have a family history of dyslexia, and
some extent, this may be because the dyslexic children are children who have a history of speech-language difficulties
older ; nonetheless, the availability of semantic context is appear to be at risk, although the nature of the speech-
an important compensatory resource for dyslexic children. language impairment is an important factor in deter-
Moreover, just as they vary in their phonological skills, mining outcome. Catts (1993) reported that children with
they will vary in their semantic abilities. At the behavioural speech-articulation problems in kindergarten had normal
level this makes for considerable variation not only in reading development at the end of first grade in contrast to
their phonological decoding skills but also in their word children who also had language difficulties. Similarly,
recognition ability (Snowling & Nation, 1997). Bishop and Adams (1990) reported that preschoolers with
specific language impairments at 4 years, whose spoken
Finally, some of the variation among dyslexics seen at the language had a good prognosis at 5" years, developed
#
behavioural level may be due to comorbidity between the normal decoding skills by 8" years, although this con-
#
disorder itself and other quite distinct disorders (Caron & clusion has to be tempered by the findings of a follow-up
Rutter, 1991). There is now a substantial body of research at age 15 showing that these children were poorer readers
suggesting that many dyslexic children have low-level than age- and language-matched peers (Stothard et al., in
impairments of the transient visual system (Hogben, press). Clearly, more research is required on the overlap
1997). Although there is no conclusive evidence that these between spoken and written language difficulties. Given

Child Psychology & Psychiatry Review Volume 3, No. 1, 1998

8
 Forum on Dyslexia : Dyslexia as a Phonological Deficit

our present state of knowledge, a tentative conclusion is Borstrøm, I., & Elbro, C. (1997). Prevention of dyslexia in
that many children with preschool speech and language kindergarten : Effects of phoneme awareness training with
impairments enjoy a reasonable start in learning to read, children of dyslexic parents. In C. Hulme & M. Snowling
perhaps because the early development of their phono- (Eds.), Dyslexia : Biology, cognition and intervention (pp.
logical awareness is supported by speech therapy. It 235–253). London : Whurr.
appears that the children at most risk of reading failure are Bradley, L., & Bryant, P. E. (1983). Categorising sounds and
those with more severe language disorders, especially learning to read—a causal connection. Nature, 301, 419–521.
where the phonological system is affected and a disorder Bruck, M. (1990). Word-recognition skills of adults with child-
of speech production is evident (Stackhouse, 1996), hood diagnoses of dyslexia. Developmental Psychology, 26,
together with those whose deficits in phonological aware- 439–454.
ness go unobserved. Bruck, M. (1992). Persistence of dyslexics’ phonological aware-
ness deficits. Developmental Psychology, 28, 874–886.
By school entry it is possible to detect, with reasonable Bryant, P.E., & Impey, L. (1986). The similarities between normal
accuracy, the children who will go on to have difficulties
readers and developmental and acquired dyslexic children.
with reading by assessing their poor performance on tests
Cognition, 24, 121–137.
of phonological awareness and letter knowledge (Muter,
Byrne B., & Fielding-Barnsley, R. (1989). Phonemic awareness
1996). With young children in the age range of 4 to 5, tests
and letter knowledge in the child’s acquisition of the alphabetic
of syllable and rhyme awareness seem the most suitable
principle. Journal of Educational Psychology, 81, 805–812.
for this purpose, whereas from 6 years onwards, tests
Byrne, B., Fielding-Barnsley, R., Ashley, L., & Larsen, K. (in
requiring awareness of spoken words at the level of the
press). Assessing the child’s contribution to reading acqui-
phoneme are the most sensitive (Muter, Hulme, &
sition : What we know and what we don’t know. In B. Blachman
Snowling, 1997). Regrettably, the most common referrals
to clinicians will be of children who have already failed to (Ed.), Cognitive and linguistic foundations of reading acquisition.
learn to read. The starting point of the assessment they Hillsdale, NJ : Lawrence Erlbaum.
undertake should involve an assessment of general cog- Caron, C., & Rutter, M. (1991). Comorbidity in child psy-
nitive ability, not least to identify the child’s strengths and chopathology : Concepts, issues and research strategies. Journal
weaknesses and to assess the specificity of their reading of Child Psychology and Psychiatry, 32, 1063–1080.
problems. However, the assessment will be most useful if Castles, A., & Coltheart, M. (1993). Varieties of developmental
it proceeds to assess the phonological skills that underlie dyslexia. Cognition, 47, 149–180.
reading development, and the reading strategies the child Catts, H. W. (1993). The relationship between speech-language
is currently using with a view to prescribing appropriate and reading disabilities. Journal of Speech and Hearing Re-
intervention (Goulandris, 1996). It is also important to search, 36, 948–958.
note that many children with reading difficulties show Coltheart, M., Masterson, J., Byng, S., Prior, M., & Riddoch, J.
signs of attention deficit disorder and hyperactivity (1983). Surface dyslexia. Quarterly Journal of Experimental
(Pennington et al., 1993). These require separate investi- Psychology, 35, 469–495.
gation and may require their own intervention. Eden, G. F., VanMeter, J. W., Rumsey, J. M., Maisog, J. M.,
Woods, R. P., & Zeffiro, T. A. (1996). Abnormal processing of
More optimistically, there is now a considerable body of visual motion in dyslexia revealed by functional brain imaging.
literature pointing to the benefits of phonological aware- Nature, 382, 66–69.
ness training to subsequent literacy development, par- Farmer, M. E., & Klein, R. M. (1995). The evidence for a
ticularly when combined with highly structured reading temporal processing deficit linked to dyslexia : A review.
tuition (Hatcher, Hulme, & Ellis, 1994 ; Snowling, 1996). Psychonomic Bulletin and Review, 2, 460–493.
However, it is also becoming clear that dyslexic children
Fowler, A. (1991). How early phonological development might
and children at risk of dyslexia (Borstrøm & Elbro, 1997 ;
set the stage for phoneme awareness. In S. A. Brady & D. P.
Byrne et al., in press ; Olson et al., in press) respond
Shankweiler (Eds.), Phonological processes in literacy : A tribute
relatively less well to these techniques than normally
to Isabelle Liberman (pp. 97–117). Hillsdale, NJ : Lawrence
developing readers. More research needs to be done to
Erlbaum.
identify the characteristics of dyslexic children that best
Frith, U. (1985). Beneath the surface of developmental dyslexia.
predict their responsiveness to such treatments. For
In K. Patterson, M. Coltheart, & J. Marshall (Eds.), Surface
children who do not respond, alternative therapies may
dyslexia (pp. 301–330). London : Routledge & Kegan Paul.
need to be designed (e.g. Wise, Olson, & Ring, 1997) and
Gallagher, A., Frith, U., & Snowling, M. (1996). Early language
special school placement considered.
skills in preschoolers at risk of dyslexia. Unpublished manu-
script, MRC Cognitive Development Unit.
References Goswami, U. (1997). Learning to read in different orthographies :
Bishop, D. V. M., & Adams, C. (1990). A prospective study of Phonological awareness, orthographic representations and
the relationship between specific language impairment, phono- dyslexia. In C. Hulme & M. Snowling (Eds.), Dyslexia :
logical disorders and reading retardation. Journal of Child Biology, cognition and intervention (pp. 131–152). London :
Psychology and Psychiatry, 31, 1027–1050. Whurr.

Child Psychology & Psychiatry Review Volume 3, No. 1, 1998

9
Forum on Dyslexia : Dyslexia as a Phonological Deficit

Gough, P. B., & Tunmer, W. E. (1986). Decoding, reading, and Paulesu, E., Frith, U., Snowling, M., Gallagher, A., Morton, J.,
reading disability. Remedial and Special Education, 7, 6–10. Frackowiak, R. S. J., & Frith, C. D. (1996). Is developmental
Goulandris, N. (1996). Assessing reading and spelling skills. In dyslexia a disconnection syndrome ? Evidence from PET
M. Snowling & J. Stackhouse (Eds.), Dyslexia, speech and scanning. Brain, 119, 143–157.
language : A practitioner’s handbook (pp. 77–107). London : Pennington, B. F. (1994). Genetics of learning disabilities.
Whurr. Journal of Child Neurology, 10 (Supplement), S69–S76.
Hatcher, P. J., Hulme, C., & Ellis, A. W. (1994). Ameliorating Pennington, B. F., Grossier, D., & Welsh, M. C. (1993). Con-
early reading failure by integrating the teaching of phono- trasting cognitive deficits in attention deficit hyperactivity
logical skills. Child Development, 65, 41–57. disorder versus reading disability. Developmental Psychology,
Hogben, J. (1997). How does a visual transient deficit affect 28, 511–523.
reading ? In C. Hulme & M. Snowling (Eds.), Dyslexia : Pennington, B. F., Van Orden, G. C., Smith, S. D., Green, P. A.,
Biology, cognition and intervention (pp. 59–71). London : & Haith, M. M. (1990). Phonological processing skills and
Whurr. deficits in adult dyslexic children. Child Development, 61,
Hulme, C., & Snowling, M. (1992). Deficits in output phonology : 1753–1778.
An explanation of reading failure ? Cognitive Neuropsychology, Rack, J. P., Snowling, M. J., & Olson, R. K. (1992). The nonword
9, 47–72. reading deficit in dyslexia : A review. Reading Research Quar-
Lefly, D. L., & Pennington, B. F. (1996). Longitudinal study of terly, 27, 29–53.
children at high family risk for dyslexia : The first two years. In Rodgers, B. (1983). The identification and prevalence of specific
M. L. Rice (Ed.), Toward a genetics of language. Hillsdale, NJ : reading retardation. British Journal of Educational Psychology,
Lawrence Erlbaum. 53, 369–373.
Lovegrove, W. J., & Williams, M. C. (1993). Visual temporal Rutter, M., & Yule, W. (1975). The concept of specific reading
processing deficits in specific reading disability. In D. M. retardation. Journal of Child Psychology and Psychiatry, 16,
Willows, R. S. Kruk, & E. Corcos (Eds.), Visual processes in 181–197.
reading and reading disabilities (pp. 311–330). Hillsdale, NJ : Scarborough, H. S. (1990). Very early language deficits in dyslexic
Lawrence Erlbaum. children. Child Development, 61, 1728–1743.
Lundberg, I. (1994). Reading difficulties can be predicted and Seymour, P. H. K. (1986). A cognitive analysis of dyslexia.
prevented. A Scandinavian perspective on phonological aware- London : Routledge.
ness and reading. In C. Hulme & M. Snowling (Eds.), Reading Shankweiler, D., & Crain, S. (1986). Language mechanisms and
development and dyslexia. London : Whurr. reading disorder : A modular approach. Cognition, 24, 139–164.
Manis, F. R., Seidenberg, M. S., Doi, L. M., McBride-Chang, Share, D. L. (1995). Phonological recoding and self-teaching :
C., & Petersen, A. (1996). On the bases of 2 subtypes of Sine qua non of reading acquisition. Cognition, 55, 151–218.
developmental dyslexia. Cognition, 58, 157–195. Shaywitz, S. E., Escobar, M. D., Shaywitz, B. A., Fletcher, J. M.,
McGee, R., Williams, S., Share, D. L., Anderson, J., & Silva, & Makuch, R. (1992). Evidence that dyslexia may represent the
P. A. (1986). The relationships between specific reading lower tail of a normal distribution of reading ability. New
retardation, general reading backwardness and behavioural England Journal of Medicine, 326, 145–150.
problems in a large sample of Dunedin boys. Journal of Child Shaywitz, S. E., Shaywitz, B. A., Fletcher, J. M., & Escobar,
Psychology and Psychiatry, 27, 597–610. M. D. (1990). Prevalence of reading disability in boys and girls :
Morgan, W. P. (1896). A case of congenital word blindness. Results of the Connecticut longitudinal study. Journal of the
British Medical Journal, 2, 1564–1570. American Medical Association, 264, 998–1002.
Muter, V. (1996). Predicting children’s reading and spelling Shaywitz, B. A., Fletcher, J. M., Holahan, J. M., & Shaywitz,
abilities. In M. Snowling & J. Stackhouse (Eds.), Dyslexia, S. E. (1992). Discrepancy compared to low achievement
speech and language : A practitioner’s handbook (pp. 31–44). definitions of reading disability : Results from the Connecticut
London : Whurr. longitudinal study. Journal of Learning Disabilities, 25,
Muter, V., Hulme, C., & Snowling, M. (1997). Manual : Phono- 639–648.
logical abilities test. London : Psychological Corporation. Silva, P. A., McGee, R., & Williams, S. (1985). Some charac-
Nation, K., & Snowling, M. (in press). Individual differences in teristics of 9-year-old boys with general reading backwardness
contextual facilitation : Evidence from dyslexia and poor or specific reading retardation. Journal of Child Psychology and
reading comprehension. Child Development. Psychiatry, 26, 407–421.
Olson, R., Wise, B., Conners, F., Rack, J., & Fulker, D. (1989). Snowling, M. (1995). Phonological processing and developmental
Specific deficits in component reading and language skills : dyslexia. Journal of Research in Reading, 18, 132–138.
Genetic and environmental influences. Journal of Learning Snowling, M. (1996). Contemporary approaches to the teaching
Disabilities, 22, 339–349. of reading. Journal of Child Psychology and Psychiatry, 37,
Olson, R., Wise, B., Johnson, M., & Ring, J. (in press). The 139–148.
etiology and remediation of phonologically based word rec- Snowling, M., Goulandris, N., & Defty, N. (in press). Individual
ognition and spelling disabilities : Are phonological deficits the and developmental differences in dyslexia. In C. Hulme & M.
‘ Hole ’ story ? In B. Blanchman (Ed.), Foundations of reading Joshi (Eds.), Reading and spelling : Development and disorders.
acquisition. Mahwah, NJ : Lawrence Erlbaum. Hillsdale, NJ : Lawrence Erlbaum.

Child Psychology & Psychiatry Review Volume 3, No. 1, 1998

10
 Forum on Dyslexia : Dyslexia as a Phonological Deficit

Snowling, M., Goulandris, N., & Stackhouse, J. (1994). Phono- Stanovich, K. E., & Siegel, L. S. (1994). The phenotypic per-
logical constraints on learning to read : Evidence from single- formance profile of reading-disabled children : A regression-
case studies of reading difficulty. In C. Hulme & M. Snowling based test of the phonological-core variable-difference model.
(Eds.), Reading development and dyslexia (pp. 86–104). Journal of Educational Psychology, 86, 1–30.
London : Whurr. Stanovich, K., Siegel, L. S., & Gottardo, A. (1997). Progress in the
Snowling, M., & Hulme, C. (1994). The development of phono- search for dyslexia subtypes. In C. Hulme & M. Snowling
logical skills. Transactions of the Royal Society B, 346, 21–28. (Eds.), Dyslexia : Biology, cognition and intervention (pp.
Snowling, M., & Nation, K. (1997). Language, phonology and 108–130). London : Whurr.
learning to read. In C. Hulme & M. Snowling (Eds.), Dyslexia : Stothard, S. E., Snowling, M. J., Bishop, D. V. M., Chipchase,
Biology, cognition and intervention (pp. 153–166). London : B., & Kaplan, C. (in press). Language impaired pre-schoolers :
A follow-up in adolescence. Journal of Speech, Hearing and
Whurr.
Language Research.
Snowling, M., Nation, K., Moxham, P., Gallagher, A., & Frith,
Swan, D., & Goswami, U. (1997). Picture naming deficits in
U. (1997). Phonological processing deficits in dyslexic students :
developmental dyslexia : The phonological representations
A preliminary account. Journal of Research in Reading, 20,
hypothesis. Brain and Language, 56, 334–353.
31–34.
Vellutino, F. R. (1979). Dyslexia : Research and theory. New
Stackhouse, J. (1996). Speech, spelling and reading : Who is at
York : MIT Press.
risk and why ? In M. Snowling & J. Stackhouse (Eds.), Dyslexia,
Wimmer, H. (1996). The early manifestation of developmental
speech and language : A practitioner’s handbook (pp. 12–30).
dyslexia : Evidence from German children. Reading and
London : Whurr.
Writing, 8, 171–188.
Stanovich, K. E. (1986a). Matthew effects in reading : Some Wise, B. W., Olson, R. K., & Ring, J. (1997). Teaching phono-
consequences of individual differences in the acquisition of logical awareness with and without the computer. In C. Hulme
reading. Reading Research Quarterly, 21, 360–407. & M. Snowling (Eds.), Dyslexia : Biology, cognition and
Stanovich, K. E. (1986b). Cognitive processes and the reading intervention (pp. 254–274). London : Whurr.
problems of learning disabled children : Evaluating the as- Yule, W., Rutter, M., Berger, M., & Thompson, J. (1974). Over-
sumption of specificity. In J. K. Torgesen & B. Y. L. Wong and under-achievement in reading : Distribution in the general
(Eds.), Psychological and educational perspectives on learning population. British Journal of Educational Psychology, 44,
disabilities (pp. 87–131). San Diego, CA : Academic Press. 1–12.

Child Psychology & Psychiatry Review Volume 3, No. 1, 1998

11